[jongalloway2]

Here's my understanding:

- Mobius is .NET Framework / Mono based and x-plat isn’t great, .NET for Apache Spark is .NET Core / .NET Standard and built with x-plat as a primary concern

- Mobius only targets up to Spark 2.0; while Spark LTS is up to 2.4 now

- .NET for Apache Spark is built to take advantage of .NET Core performance improvements, showing big advantages over Python and R bindings, especially when user defined functions are a major factor

- .NET for Apache Spark is driven by lessons learned and customer demand, including major big data users inside and outside Microsoft

Disclaimer: I know people that worked on this and helped from .NET Foundation side, but the above is my non-official summary from readme's and stuff.